Disclaimer: Pre-published articles are not considered final versions. Any changes will be reflected in the online version the month the article is officially published.
Author affiliations: University of California, Los Angeles, California, United States (L. Wang, E. Vikram, A. Zou, G. Cheng); New York-Presbyterian Queens Hospital, Flushing, New York, USA (D. Liu, L. Yung, GD Rodriguez, N. Prasad, S. Segal-Maurer, V. Singh, WH Rodgers); Weil Cornell Medical College, New York (WH Rodgers)
Mycobacteria are major human pathogens; â13 million people in the United States live with Mycobacterium tuberculosis complex (MTBC) and the incidence of non-tuberculous mycobacterial lung disease (NTM) is increasing worldwide. The challenges of managing MTBC and M. avium complex coinfections (CAM) are well described, including the risk of misinterpreted Xpert RIF (rifampin) results (1,2). MTBC and Mr. abscess co-infection has only been described in case reports (3,4). We describe the co-infection with 4 species of mycobacteria.
In July 2019, an 82-year-old Asian woman was hospitalized in Flushing, New York, United States, with a persistent fever associated with worsening weakness. A CT scan of her chest showed almost complete left upper lobe atelectasis, hyperinflation in other areas, and a small left pleural effusion. Scattered bud tree-shaped nodular opacities and pulmonary granulomas were present in all lungs, and discontinuity of the left upper lobe bronchus was noted. Cultures of blood, urine, stool and respiratory samples gave negative results for non-mycobacteria.
In a sputum sample collected for routine mycobacterial testing, fluorochrome staining showed rare acid-fast bacilli, and MTBC was detected using Xpert MTB / RIF (Cepheid, https: //www.cepheid. com). We then inoculated sputum at a Lowenstein-Jensen Gruft slope, incubated it at 37 Â° C, and inoculated VersaTREK Myco bottles containing Middlebrook 7H9 broth modified with sponges (Thermo Fisher, https://www.thermofisher.com ) and incubated them at 35 Â° C. No isolates were recovered from the Lowenstein-Jensen Gruft slope. Only MAC was detected by AccuProbe (Hologic, https://www.hologic.com) in the positive Kinyoun culture of Myco bottles. A week later, another sputum sample with growth positive for Kinyoun from the Myco vials was negative for MAC, MTBC, Mr. gordonae, and Mr. kansasi. Mr. abscess was identified on the Lowenstein-Jensen Gruft slope.
Given the sensitivity limit and narrow species coverage of AccuProbe and the difficulty of identifying mycobacteria by culture and due to growth interferences between different mycobacteria, we performed next-generation sequencing (NGS ) using Hiseq3000 (Illumina, https: //www.illumina. com) on the first sputum culture supernatant. NGS generated â 175 million reads, each with a quality score> 35. We checked the NGS data for quality control using FastQC (Galaxy, https://usegalaxy.org). All stages and programs used Galaxy’s data processing pipeline, an open-source web-based platform for data-intensive biomedical research. Each identified reading had a quality control score of 39.4 and an average guanine-cytosine content of 68%. Only 0.69% of the bases gave no results and were not identifiable. We performed De Novo classification using De Novo Assembly Unicycler, Quast QC and Kraken Classification (Galaxy) and generated coverage and depth data using BWA Aligner and SAMtools Depth (Galaxy). We aligned the reads, visualized on the bacteria’s reference genomes using Bowtie2 (Galaxy) and converted to BED (Browser Extensible Data) files followed by Bedtools Merge, Multicov (https://bedtools.readthedocs.io) .
The genome visualization pipeline confirmed 4 genomic traces of Mycobacterium strains (Figure): Mr. yongonense strain 05-1390 (GenBank accession number NC_021715.1), M. tuberculosis strain FDAARGOS_757 (GenBank accession number CP054013.1), Mycobacterium sp. MOTT36Y (GenBank accession number NC_017904.1), and M. abscess ATCC 19977 (GenBank accession number CU458896.1). Mr. yongonense was identified with a genomic coverage of 88.73% (4.9 Mb mapped on 5.5 Mb genome) and a read depth of 1224 Ã. M. tuberculosis was identified with a genomic coverage of 99.99% (4.4 Mb mapped onto 4.4 Mb genome) and a read depth of 63 Ã. Mycobacterium sp. was identified with a genome coverage of 94.41% (5.3 Mb mapped from the 5.6 Mb genome) and a read depth of 1210 Ã. Mr. abscess was identified with a genome coverage of only 2.75% (0.14 Mb mapped from the 5.1 Mb genome) and a read depth of 8 Ã (table). The mycobacteria identified by NGS were verified by various mycobacteria tests.
We obtained the consensus sequence for 4 strains of bacteria using MEGAHIT (Galaxy) and generated a BLAST tree (https://blast.ncbi.nlm.nih.gov/Blast.cgi) based on minimal evolution at the level of the species using >15 kbp of each sequence. Assembly on MTBC sequencing data gave a total consensus sequence of 4,376,826 bp and 78,208 single nucleotide polymorphism sites (1.79%). Analysis by BLAST and Mykrobe (https://www.mykrobe.com) revealed that the MTBC isolate belongs to the 2.2 subline.
The patient received RIPE treatment (rifampin, isoniazid, pyrazinamide and ethambutol), as well as amikacin, tigecycline and azithromycin. At 6 months, the RIPE treatment was finished. At 9 months, the sputum culture was negative. The patient continues to take amikacin, tigecycline and azithromycin on an outpatient basis with close monitoring.
Identification of co-infection with mycobacteria is necessary for diagnosis and treatment (5). Treatment regimens and duration remain species specific due to unique resistance mechanisms. To achieve the greatest potential for success while minimizing toxicities, early empiric treatment should take into account the clinical features of MTBC and NTM co-infection and strain identification.
Our report highlights the value of NGS in identifying multiple mycobacterial co-infections in populations with high susceptibility and prevalence of MTB and NCD (i.e. immigrants, immunocompromised patients and international travelers) . NGS can trace the geographic origin of the Mycobacterium stump. These characteristics, combined with a patient’s epidemiological exposure and travel history, could elucidate the potential time and place of infection acquisition. NGS could also be used to identify drug resistance genes to guide targeted therapy.
Dr Wang is a postdoctoral researcher in the Department of Microbiology, Immunology and Molecular Genetics at the University of California, Los Angeles. His research focuses on immunology, virology and molecular diagnostics. Dr Liu is the Director of the Laboratory of Microbiology, Immunology and Molecular Diagnostics, Department of Pathology and Clinical Laboratories, New York-Presbyterian Queens Hospital. His research focuses on the development of rapid molecular tests for the diagnosis of pathogens.
The conclusions, findings, and opinions expressed by the authors contributing to this review do not necessarily reflect the official position of the United States Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or institutions affiliated with the. authors. . Use of trade names is for identification purposes only and does not imply endorsement of any of the groups mentioned above.